As of 13 August 2026, AI can only partly find your missing expense receipts.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Export the relevant business bank or card transactions as a CSV or spreadsheet, including the date, supplier, amount, currency and payment reference.
   2. Open the receipt storage locations you are authorised to use, such as the expense system, shared drive, cloud folder and work email, then create a file index containing each receipt filename, date, supplier and amount without copying unnecessary personal data.
   3. Paste the transaction export and receipt index into the prompt, keeping the column headings and identifying which list is which.
   4. Ask the model to produce the matching table, then save the rows marked missing receipt, possible match or insufficient data.
   5. Search the suggested supplier names and date ranges in your work email, cloud folders and paper records, and attach any receipts you find to the corresponding expense record.
   6. Compare every remaining flagged transaction with the original bank or card statement and the receipt itself, then ask the person responsible for the expense or supplier for a replacement receipt where appropriate.
   7. Give the final unresolved list to your bookkeeper or accountant before claiming VAT or submitting the expenses, and retain the source records with the completed expense file.

   Prompt

   I need to find missing expense receipts for a UK business. I will provide an expense transaction list and a separate receipt register or folder index. Compare them using the transaction date, supplier, amount, currency and payment reference where available. Do not invent receipts, suppliers, dates, amounts or matches. Treat small date differences, card settlement dates and VAT-inclusive versus VAT-exclusive amounts as possible reasons for a near match, not as proof of a match.
   
   Return a table with these columns: transaction date, supplier, amount, currency, payment reference, matching receipt or file, match status, confidence, and reason. Use only these match statuses: matched, probable match, possible match, missing receipt, or insufficient data. Put every uncertain item in a separate review list. Also list duplicate receipts, receipts whose details do not agree with the transaction, and transactions that could be personal or otherwise unsupported. Do not decide whether an expense is allowable for tax or VAT, and do not give professional advice. Finish with a short list of searches I should make in email, cloud storage and paper files, using the supplier names and date ranges found in the data.
   
   Here is the expense transaction list:
   [PASTE TRANSACTION LIST]
   
   Here is the receipt register or file index:
   [PASTE RECEIPT REGISTER OR FILE INDEX]

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.